This is covered in the tutorial for GKE here.

Firstly, you should reserve a static IP address, so that it is re-used between deployments.

You can do this in the GCP console or via the command, like so: gcloud compute addresses create helloweb-ip --region us-central1

Then you can find the static IP address you have been allocated: gcloud compute addresses describe helloweb-ip --region us-central1

Then you should use this IP address when setting up your LoadBalancer service. In "imperative" commands, you would type:

kubectl expose deployment my-app --target-port=443 --port=443 --type=LoadBalancer --loadBalancerIP=YOUR.IP.ADDRESS.HERE

The YAML "declarative" form is convenient as the setup gets more complex, this can also saved for re-use in future. For example:

apiVersion: v1
kind: Service
metadata:
  name: helloweb
  labels:
    app: hello
spec:
  selector:
    app: hello
    tier: web
  ports:
  - port: 80
    targetPort: 8080
  type: LoadBalancer
  loadBalancerIP: "YOUR.IP.ADDRESS.HERE"

Yes, you can use already existing service to point new deployment.

All you've to do, edit the service YAML with newly deployed service's label.

Example

Suppose, you've newly deployed services yaml as below.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
  labels:
    app: nginx
spec:
  replicas: 3
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: my-app-2
    spec:
      containers:
      - name: nginx
        image: nginx:1.7.9
        ports:
        - containerPort: 80

It has label as

template:
    metadata:
      labels:
        app: my-app-2

If your service is as below:

kind: Service
apiVersion: v1
metadata:
  name: my-app
spec:
  type: LoadBalancer
  selector:
    app: my-app
  ports:
  - protocol: TCP
    port: 80
    targetPort: 80

you only need to change selector section of service

selector:
    app: my-app

to

selector:
    app: my-app-2